A local-first, offline PWA for composing LLM prompts from reusable text snippets.
The core loop: select an Agent > activate Snippets > copy the compiled output.
When you first open the app, an onboard wizard will appear. You can always retrigger this wizard by opening up the Settings menu and clicking Restart Tour.
| Document | Summary |
|---|---|
| Architecture | App layers, tech stack, and folder structure |
| Contributing | Setup, branch workflow, and code quality gates |
| Models | Data model — Agents, Skills, Snippets — entity types, fields, cascade rules |
| Spec Kit | How to use spec-kit with this project |
| State and Data Flow | Store, compiler, persistence, and import/export |
| Styling | Design philosophy, component libraries, motion libraries |
pn install # Install dependencies with pnpm
pn dev # Start dev server with HMR (http://localhost:5173)
pn lint-all # Lint and format the code
pn test-all # Run unit and e2e tests
pn build # Build the app
pn commit # Create styled commit messages